Did I waste time getting an MS in computational physics?

[e15]

I completed the first year of grad school in physics and then contemplated returning to get the master's or just drop out. I had some prior experience using C++ and C# and was getting interviews for C# developer roles, but could not pass technical interviews since I failed questions on data structures/algorithms, knowing abstract classes/interfaces, etc. Since I was more interested in getting a job as a data scientist or quant or some role using more math than software engineering, I was told that I was better off finishing the master's in physics than switching to stats

This past year I took courses in numerical linear algebra, Bayesian statistical methods, and using Matlab for numerical methods. I did my master's thesis on a computational physics project using C++ and unix/linux. After months of applying, I still can't get an offer for a software engineer, quant, or data scientist role. I've been told that because I lack work experience and experience working on large data sets, I'm not qualified for data scientist roles. I also haven't been able to pass the coding tests for software engineer roles.

I feel like I was better off spending this past year just studying C# coding problems and getting a C# developer role than completing the Master's. Am I right? Is there anything I can do to get a data scientist role other than just working on problems on kaggle?

 

[jedishrfu]

Since you're getting past the first hurdle and getting to the test part then you need to maybe get certified in some programming language and that should prepare you for these kinds of tests. I think that's what's stopping you from getting hired.

I'm sure there are smaller companies around that won't do this. I encountered this only once in my long career when I was interviewed by a junior programmer who didn't know how to interview and instead asked some basic comp 101 type questions like show me how you'd swap two values x and y in code.

 

[Vanadium 50]

As I understand what you wrote, you needed to learn more about data structures to achieve your goals, but you went down a path that didn't do this. Is this right? If so, yes, I think you wasted your time.

 

[Jon321]

What is your bachelor's degree in? What location(s) are you applying to jobs in? Which school did you do your MS at? Many data science jobs want a PhD, so perhaps that is part of the problem. Anyway, it's a little surprising to me because there is a huge demand for people with quantitative data analytics skills.

 

[e15]

My BS is in physics. I've applied to jobs at many places, including the bay area, NY, Chicago, etc. I am completing my MS at a large state school.

 

[jedishrfu]

V50 is right for many CS jobs you will need data structures and algorithms skills as well as web application and database skills. For large data you will need skills in statistics, data mining and database design.

 

[Vanadium 50]

for many CS jobs you will need data structures and algorithms skills

That's true, but I think my point is even stronger - if you apply for jobs, and the common theme on why they don't hire you is that you need to know more about data structures, I think many people - especially the more successful ones - would then focus on data structures.

 

[Jon321]

My BS is in physics. I've applied to jobs at many places, including the bay area, NY, Chicago, etc. I am completing my MS at a large state school.

Hum, its a bit surprising to me that you haven't had any offers, assuming your GPA is OK (>3.25). Here are some ideas:

1. Try applying to some lesser known companies or startups, including companies that pay just as well but are less sexy. Today, companies like Amazon, Twitter, Google and Pinterest get incredible numbers of applications. On the other hand, there are many jobs that pay as well or better but are less "hip", like big retail, pharmaceutical, agricultural, industrial, financial or oil/gas/mining companies, government R&D labs.
2. At the MS level, you might be competing more with software engineers than "data scientists". Therefore, you should take the Coursera Algorithms I and II courses offered by Stanford. They are free at Coursera and taught by Tim Roughgarden. Also buy the course book. I think this will teach you enough.
3. C++ isn't used very much outside of finance and defense contractors, and Matlab isn't used that much outside of engineering circles. Its good to know C++, but you should also know Java and probably R.

Assuming your GPA is OK, it is virtually impossible that you "wasted your time" getting an MS in computational physics. But you do have to understand that such a degree, by itself, doesn't prepare you for the typical mobile app / website / SaaS software engineer jobs that are in vogue today. Instead, you need to find jobs that actually require knowledge of physics, probability and numerical methods. Those are not going to be at the next $50 billion social networking startup. And I wouldn't be surprised if a lot of the jobs being advertised as "data scientist" jobs are really just database administrator or software engineer jobs. Unfortunately, 99% of US industry these days does not want to spend 2 cents on any R&D or other efforts that require the intelligence and deep thought abilities of a physicist. Instead that money goes to stock buy backs, nowadays. That's why physicists take jobs on Wall Street.

 

[vela]

That's true, but I think my point is even stronger - if you apply for jobs, and the common theme on why they don't hire you is that you need to know more about data structures, I think many people - especially the more successful ones - would then focus on data structures.

That was the common theme for the developer jobs e15 didn't particularly want. If you have to choose one or the other, why focus on developing skills for a job you'd have to settle for rather than for the career you want?

 

[e15]

Hum, its a bit surprising to me that you haven't had any offers, assuming your GPA is OK (>3.25).

My gpa is around 3.4. I left it off my resume since that's considered poor for grad school, but perhaps it looks even worse if I leave it off?

Here are some ideas:

1. Try applying to some lesser known companies or startups, including companies that pay just as well but are less sexy. Today, companies like Amazon, Twitter, Google and Pinterest get incredible numbers of applications. On the other hand, there are many jobs that pay as well or better but are less "hip", like big retail, pharmaceutical, agricultural, industrial, financial or oil/gas/mining companies, government R&D labs.
2. At the MS level, you might be competing more with software engineers than "data scientists". Therefore, you should take the Coursera Algorithms I and II courses offered by Stanford. They are free at Coursera and taught by Tim Roughgarden. Also buy the course book. I think this will teach you enough.
3. C++ isn't used very much outside of finance and defense contractors, and Matlab isn't used that much outside of engineering circles. Its good to know C++, but you should also know Java and probably R

I've already applied to many big retail, financial and oil/gas companies. However, I haven't applied to that many startups

I also already know R. I have a few months of experience using it on course and personal projects

Instead, you need to find jobs that actually require knowledge of physics, probability and numerical methods.

I've seen very few jobs that require knowledge of numerical methods (except jobs requiring a security clearance to work on HPC. Although I'm interested in HPC, I don't want to get a clearance)

 

[Jon321]

Personally, I would be more suspicious of somebody who left their GPA off of their resume than I would if they put down a 3.4. If somebody doesn't put down a GPA, I'd assume its downright horrible (<3.0). A lot of MS level courses are still graded such that they are B- centered, whereas a lot of PhD courses are more A- centered.

Also, if you rule out jobs that require a security clearance, you are basically ruling out all jobs that would actually use the skills of a computational physicist (aside from low paying academic research assistant jobs).